Receive as a parameter the instrument in the rule function

[diegolovison]

I would like to receive the instrument into the rule function.
Is it possible?

    def my_rule(self, price):
        signals = pd.Series(dtype='float64', index=price.index)
        for index in price.index:
            signal_date = index.replace(day=1)
            if signal_date in self.signals.index:
                signals.loc[index] = self.signals[instrument][signal_date]
        return signals

    def run(self):
        my_rule = TradingRule(self.my_rule)
        my_rules = Rules(dict(my_rule=my_rule))
        fcs = ForecastScaleCap()
        my_system = System([fcs, my_rules], self.data)
        print(my_system.forecastScaleCap.get_forecast_scalar("PETR4", "my_rule").tail(5))

[diegolovison]

How about modify the code to. I could fill the PR
trading_rules.py

        list_of_data_for_call = self._get_data_from_system(system, instrument_code)
        result = self._call_with_data(list_of_data_for_call, instrument_code)

        # Check for all zeros
        result = replace_all_zeros_with_nan(result)

    def _call_with_data(self, list_of_data_for_call: list, instrument_code: str) -> pd.Series:
        other_args = self.other_args
        other_args['__instrument_code'] = instrument_code
        result = self.function(*list_of_data_for_call, **other_args)
        return result

[robcarver17]

This shouldn't be neccessary. I don't understand the use case why you would want to do that? The whole poiint is the system prepares the data for a specific rule, and passes it to the forecasting function.

[diegolovison]

The "my_rule" method returns the raw_forecast based on price. In my case, to generate the raw_forecast, I am using an external system. Without knowing the instrument, I am not able to request a raw_forecast.

Example:

    def my_rule(self, price):
        instrument_code = ....
        signals = pd.Series(dtype='float64', index=price.index)
        for index in price.index:
            signal_date = index.replace(day=1)
            if signal_date in self.signals.index:
                signals.loc[index] = self.signals[instrument_code][signal_date]
        return signals

        my_rule = TradingRule(self.my_rule)
        my_rules = Rules(dict(my_rule=my_rule))
        fcs = ForecastScaleCap()
        my_config = Config()
        accounts = Account()
        position_size = PositionSizing()
        raw_data = RawData()
        portfolio = Portfolios()
        combiner = ForecastCombine()
        my_system = System([fcs, my_rules, combiner, raw_data, portfolio, position_size, accounts], self.data, my_config)

[robcarver17]

You should create a new method in raw data that pulls in your external stuff.

…

On Fri, 21 Oct 2022, 23:43 Diego Lovison, ***@***.***> wrote: The "my_rule" method returns the raw_forecast based on price. In my case, to generate the raw_forecast, I am using an external system. Without knowing the instrument, I am not able to request a raw_forecast. Example: def my_rule(self, price): instrument_code = .... signals = pd.Series(dtype='float64', index=price.index) for index in price.index: signal_date = index.replace(day=1) if signal_date in self.signals.index: signals.loc[index] = self.signals[instrument_code][signal_date] return signals my_rule = TradingRule(self.my_rule) my_rules = Rules(dict(my_rule=my_rule)) fcs = ForecastScaleCap() my_config = Config() accounts = Account() position_size = PositionSizing() raw_data = RawData() portfolio = Portfolios() combiner = ForecastCombine() my_system = System([fcs, my_rules, combiner, raw_data, portfolio, position_size, accounts], self.data, my_config) [image: Capturar] <https://user-images.githubusercontent.com/687311/197298924-4b61a058-cc42-43f5-9817-0a3955a7eb3e.PNG> — Reply to this email directly, view it on GitHub <#808 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADCLPEULO7PIQD5VRYOOVSDWEMMCDANCNFSM6AAAAAARKUEA2U> . You are receiving this because you commented.Message ID: ***@***.*** com>

[diegolovison]

I didn't understood your idea.

Then, I tried the following

I can extends TradingRule and overwrite _call_with_data.
In this case, no changes in pysystemtrade
Do you think it is a good idea?

class MyTradingRule(TradingRule):

    def _call_with_data(self, list_of_data_for_call: list, instrument_code: str) -> pd.Series:
        other_args = self.other_args
        other_args['__instrument_code'] = instrument_code
        result = self.function(*list_of_data_for_call, **other_args)
        return result

[robcarver17]

This is totally unnecessary, and I am pretty sure means that you'd break all the existing trading rules.

You really need to read the documentation (again?).

import pandas as pd
from systems.provided.futures_chapter15.basesystem import *
class newRawData(RawData):
    def get_some_data_for_instrument(self, instrument_code: str) -> pd.Series:
        ## could come from anywhere!
        return pd.Series(3.141, self.data_stage[instrument_code].index)
def rule(data_provided):
    return data_provided
trading_rules = Rules(dict(rule1=(rule, 'rawdata.get_some_data_for_instrument', {})))
system = System([newRawData(), trading_rules], data = csvFuturesSimData())
print(system.rules.get_raw_forecast('US10','rule1'))

[diegolovison]

It worked. hmm, I wasn't able to find that info inside docs folder. anyway, thanks for your time